Conducting qualitative interviews by telephone: Lessons learned from a study of alcohol use among sexual minority and heterosexual women

Abstract

This study explored effective interviewer strategies and lessons-learned based on collection of narrative data by telephone with a sub-sample of women from a population-based survey, which included sexual minority women. Qualitative follow-up, in-depth life history interviews were conducted over the telephone with 48 women who had participated in the 2009–2010 National Alcohol Survey. Questions explored the lives and experiences of women, including use of alcohol and drugs, social relationships, identity, and past traumatic experiences. Strategies for success in interviews emerged in three overarching areas: 1) cultivating rapport and maintaining connection, 2) demonstrating responsiveness to interviewee content, concerns, and 3) communicating regard for the interviewee and her contribution. Findings underscore both the viability and value of telephone interviews as a method for collecting rich narrative data on sensitive subjects among women, including women who may be marginalized.

Introduction

Although the use of telephones for collecting quantitative survey data is common and well-represented in research literature, using telephones for qualitative interviews has generally been considered an inferior alternative to face-to-face interviews (Novick, 2008). Qualitative interviews provide “a way of generating empirical data about the social world by asking people to talk about their lives” (Holstein and Gubrium, 2003, p. 3). Consequently, concerns about the use of telephones for qualitative interviews described in the literature predominately focus on the possible negative impact on the richness and quality of empirical data collected by telephone compared to face-to-face interviews (Novick, 2008; Irvine et al., 2013). Some of the most commonly articulated concerns about telephone interviews include the challenges to establishing rapport, the inability to respond to visual cues, and potential loss of contextual data (i.e., the ability to observe the individual in a work or home environment) (Novick, 2008; Holt, 2010; Smith, 2005).

A small but growing body of literature has documented the potential of in-depth telephone interviews as a viable option for qualitative research. Many methodological studies point to logistical conveniences and other practical advantages of telephone interviews, including enhanced access to geographically dispersed interviewees, reduced costs, increased interviewer safety, and greater flexibility for scheduling (Cachia and Millward, 2011; Sturges and Hanrahan, 2004; Shuy, 2003; Carr and Worth, 2001; Musselwhite et al., 2007; Stephens, 2007). In addition to benefits related to convenience, several studies emphasize the methodological strengths of conducting qualitative interviews by telephone, such as perceived anonymity, increased privacy for respondents, and reduced distraction (for interviewees) or self-consciousness (for interviewers) when interviewers take notes during interviews (Cachia and Millward, 2011; Sweet, 2002; Stephens, 2007; Lechuga, 2012). For example, privacy was noted as a strength of telephone interviews as an alternative to interviewing patients in a hospital setting (Carr and Worth, 2001) or interviewing family members of incarcerated adults in a county jail (Sturges and Hanrahan, 2004). Telephone interviews may also require interviewees to be explicit in follow-up questions, rather than relying on non-verbal cues (Cachia and Millward, 2011). Finally, some researchers suggest that telephone interviews in qualitative data collection may mediate power dynamics that might otherwise emerge in the researcher-subject relationship. Specifically, telephone interviews, compared to in-person interviews, may be less intrusive and confer greater power and control to interviewees in terms of negotiating interviews to suit their schedules as well as rescheduling interrupting, or ending the interview (Holt, 2010; Trier-Bieniek, 2012; Saura and Balsas, 2014).

A few studies have compared explicitly the outcomes and dynamics between telephone and in-person qualitative interviews. Irvine and colleagues found that interviews by telephone were somewhat shorter in duration than in-person interviews (Irvine, 2011; Irvine et al., 2013). By contrast, other studies found no differences in the length, depth, and type of responses between telephone and in-person interviews (Sturges and Hanrahan, 2004; Sweet, 2002; Vogl, 2013). In one of the few studies to examine dynamics between interviewer and interviewee, Irvine and colleagues (2013) found that interviewees were more likely to make requests for clarification and to check on the adequacy (specifically the sufficiency and relevance) of their responses during telephone interviews (Irvine et al., 2013). Although studies typically have not investigated interviewee perspectives on their experiences, Sturges and Hanrahan (2004) allowed interviewees to select their mode of interview and followed up with a question about their perception of the choice at the close of the interview. Interviewees were equally positive about their respective interview modes, but telephone interviewees were more likely to comment on privacy as an advantage (Sturges and Hanrahan, 2004).

In addition to the growing body of literature documenting telephone interviews as a viable mode for collecting qualitative data, a few studies draw on case examples to focus more explicitly on “how-to” strategies for successful in-depth telephone interviewing (Glogowska et al., 2011; Musselwhite et al., 2007; Smith, 2005). In general, these studies emphasize the importance of ensuring that advance communications (such as letters) and initial telephone communications (such as telephone scripts) communicate the purpose of the study and the importance of the participant contribution (Glogowska et al., 2011; Musselwhite et al., 2007; Smith, 2005). Other important success strategies include scheduling interviews at times convenient to interviewees, establishing rapport through small-talk and reviewing the purpose of the interview (Glogowska et al., 2011), and taking time for pre-interview training and post-interview de-briefing (Glogowska et al., 2011; Smith, 2005). Several authors of methodological studies of qualitative telephone interviews caution explicitly against “cold-calls” (Glogowska et al., 2011; Musselwhite et al., 2007; Smith, 2005). For example, to avoid cold-calls, some studies described establishing rapport and recruiting participants in person, then scheduling a telephone interview at a later date (Sweet, 2002; Sturges and Hanrahan, 2004; Musselwhite et al., 2007).

Irvine and colleagues (2013) note that there is still a need for methodological studies on qualitative telephone interviews focused on different topics and with interviewees who have different background characteristics. Methodological studies on the population-based National Alcohol Survey in the United States suggest that telephone interviews, like face-to-face interviews, are effective in collecting data about alcohol use, alcohol problems, and other sensitive questions in population-based surveys (Greenfield et al., 2000; Midanik and Greenfield, 2003; Midanik et al., 2001). However, few studies explore the utility and strategies associated with collecting narrative data by telephone on sensitive topics such as alcohol use, sexuality, and traumatic life-experiences. Because telephone interviews appear to increase a perceived sense of anonymity (Greenfield et al., 2000; Sturges and Hanrahan, 2004), they may be well suited to collecting data on sensitive topics (Trier-Bieniek, 2012), but additional studies are needed to verify whether such interviews yield rich data.

In addition, there is a paucity of studies examining the effectiveness of telephone interviews with stigmatized or hard-to-reach populations, an issue that is particularly salient to social work, health, and other applied sciences. One notable exception is the study by Sturges and Hanrahan (2004), which involved collecting qualitative data from correctional officers and visitors at county jails. The authors noted that “visitors to the jail felt stigmatized by their association with a jail inmate” (p. 114) and that the relative anonymity of the telephone interview reduced anxiety about participating in an interview. Sturges and Hanrahan also emphasized that recruitment of all participants was accomplished face-to-face, and that it was unclear whether another method of sampling would have been effective with hard-to-reach respondents. Another exception is a study by Trier-Bieniek (2012), who interviewed women over the telephone about their emotional connection to an artist whose music addressed themes such as sexual violence, religious upbringing, and repressed sexuality. Trier-Bieniek found that use of the telephone appeared to create a safe space for participants to share traumatic experiences because of the relative anonymity and opportunity to stay in settings that were comfortable to them during phone interviews. The question of whether telephone interviews are viable with stigmatized or marginalized populations when using recruiting methods other than face-to-face (e.g., Sturges and Hanrahan) or internet social networks (e.g., Trier-Bieniek) is relatively unexplored.

The current study is based on our experience conducting in-depth follow-up interviews with women participants of the population-based National Alcohol Survey, including an oversample of sexual minority women. The aim of this study was twofold: 1) to explore effective interviewer strategies for collection of narrative data on sensitive topics by telephone, and 2) to document challenges and advantages of telephone interviews with a population-based sample, including marginalized populations of women such as sexual minority women.

Methods

Research Design

This research was conducted as part of a larger study examining mediators of hazardous drinking among sexual minority women compared to heterosexual women based on the National Alcohol Survey (NAS), a national telephone-based quantitative household probability survey. The term “sexual minority” is increasingly used in research to describe diverse groups who identify as lesbian, bisexual, or whose behavior or attractions are other than exclusively heterosexual (Mayer et al., 2008); we use this term in the current study while noting that it does not reflect the variation in language that individual respondents used to describe their own identities during the interviews. Qualitative in-depth life history interviews were conducted over the telephone with women who were originally recruited as a part of a sample from the 2009–2010 NAS (n=3,825 women). The original quantitative interviews were conducted from 12–24 months prior to the qualitative interview after separate funding was acquired for this study of sexual minority women.

Sample

The original NAS probability sample included women who were classified as sexual minorities (identified as lesbian or bisexual and heterosexual-identified women who reported same-sex partners, n=122). The sampling frame for the follow-up interviews, which was designed to generate an oversample of sexual minority women, included all sexual minority women and a matched sample of heterosexual women. The matched heterosexual sample was created by generating list of randomly selected exclusively heterosexual women matched to key characteristics including age, ethnicity, relationship status, education, drinking status in the past year (drinker/none), and a lifetime measure of having consumed five or more drinks at least monthly throughout at least one decade of their life. The lifetime measure was constructed as a dichotomous variable based on responses to questions about how often respondents had five or more drinks on one or more occasions in each decade of their life (i.e., teens, 20’s, 30’s, and 40’s). The lifetime five-plus measure correlated with lifetime alcohol problem and dependence measures, and was assessed as a useful variable for matching respondents with a history of heavier drinking. The list of prospective heterosexual matches was identified as interviews with sexual minority women progressed. We did not plan to have one-to-one matching and this process allowed us to obtain matches for individuals or groups of respondents who shared similar characteristics. For example, one white heterosexual women, aged 50–59, in a partnered relationship and with a high school education, might serve as “match” for three sexual minority women with similar characteristics.

Excluding disconnected/wrong numbers or ineligible respondents, the response rate was 50 percent (48 interviewed, 26 refusals, 5 incomplete interviews, and 17 no response). Approximately 27.9 percent (n=41) were wrong numbers or no longer operative. Two of the individuals initially contacted were men (and not eligible for participation), and seven were mono-lingual Spanish-speaking (interviews were conducted in English). A 50% response rate has been typical for U.S. telephone surveys since the widespread use of caller identification (Keeter et al., 2006). There were no significant differences in final response rates between prospective interviewees who were mailed information in advance compared to those who were contacted solely by telephone.

To assess for nonresponse bias, we conducted several analyses to compare the final interview sample to non-respondents (including 17 sexual minority or heterosexual women who did not respond and 41 wrong or disconnected numbers; n=63) and refusals (who were contacted but declined; n=26). There were no differences in drinking measures (abstaining, drinking, heavier drinking, dependence, alcohol-related consequences, or past treatment) and few demographic differences between the final sample and non-respondents or refusals. No differences were found by ethnicity or relationship status. Non-respondents were generally younger and less educated than interviewees, and refusals were less likely to be employed.

The final sample included 32 sexual minority women (15 lesbians, 10 bisexuals, and 7 women who identified as heterosexual and reported same-sex partners) and 16 matched exclusively heterosexual women. Age of participants ranged from 21 to 67 years of age. Approximately 64.6 percent (n=31) of the participants were White, 22.9 percent (n=11) were African American, and 12.5 percent (n=6) were Latina. Approximately 31.3 percent (n=15) were heavier drinkers at some point in their lives. The original NAS study had an oversample of Blacks and Hispanics thus resulting in slightly disproportionate numbers of minorities, which was considered an advantage. Additional details about the follow-up interview methods are available in published manuscripts about a pilot study (Condit et al., 2011) and analysis of narratives (Drabble and Trocki, 2013).

Procedures

Prior to initiating telephone calls, a number of efforts were made to facilitate ease of communication with prospective interviewees. Approximately half of the sexual minority women in the original NAS provided a contact address and were mailed notification in advance of the initial telephone contact. The research team created a web page with information about the NAS and the follow-up study. The address of the web page was provided to prospective interviewees in the advance letter. A toll-free number was also established to minimize potential cost barriers for participants in calling with questions or to schedule an interview.

Three members of the research team conducted interviews: one lead investigator with prior experience conducting research based on qualitative interviews and two MSW graduate students. Graduate student researchers received training on skills, attributes, practices, and specific project tools for coordinating and conducting interviews. Students also reviewed pretest interview transcripts; they participated in role-plays; in addition, they discussed personal strengths, experiences and possible preconceived perceptions that they might bring to the research project and the interviews. In this process, students were trained in establishing rapport, minimizing interviewer bias, using probing questions, managing transitions, and determining when they had sufficient information to move to the next question. The senior research team member and students also debriefed after interview sessions to discuss dynamics of the interview, reflect on personal reactions, examine what went well, and identify opportunities for improving future interviews. Preparation for addressing possible distress among respondents included the adoption of a detailed protocol which provided interviewers with tools to recognize and respond to possible signs of distress, identification of local referral resources in advance (based on phone exchanges or zip code), and back-up support from senior research team members (on site) and clinical psychologists (on call) during interview sessions.

Interviews were conducted between March and December of 2011. A semi-structured interview guide was used, which included eight primary questions and follow-up probes related to study participants’ life experiences in several areas including family, friendships, identity, substance use, intimate relationships, trauma, and management of mood. Interviews generally ranged from approximately 45 minutes to 1.5 hours (mean 62.5 minutes). There were no differences in length of interviews between sexual minority and exclusively heterosexual respondents. Each respondent was given a $25 gift card. Interviews were audio taped and transcribed.

Analysis

For the purposes of this paper, a content analysis of all 48 interviews was conducted to explore strategies employed by interviewers to engage interviewees and obtain rich narrative data. Qualitative data were managed with the assistance of a qualitative software program (NVIVO). Initial open coding to conceptualize, compare, and categorize data was followed by an iterative process to further define and identify connections between categories in the data (Strauss, 1987; Strauss & Corbin, 1990). Specifically, two authors read a sample of transcripts and independently developed a list of provisional codes, which were reviewed and refined to use for coding transcripts. This was followed by an iterative process to further refine and use codes for defining and categorizing elements of successful interviews. For the purposes of analysis, success was defined as engagement of respondents from initial contact and throughout the lengthy interview, and respondent willingness to disclose and share their perceptions and experiences in multiple life domains, including personal or potentially sensitive topics. The authors used a consensus model in reviewing, revising, and finalizing categories, as well as in identifying connections between categories to define overarching themes. In addition, a separate review of debriefing notes was conducted to identify “lessons learned” from both planning and implementation phases of the interviews.

Results

Effective Interviewer Strategies: Rapport, Responsiveness, and Regard

Specific strategies for success in interviews emerged in three overarching thematic areas: 1) cultivating rapport and maintaining connection, 2) demonstrating responsiveness to interviewee content, concerns, and 3) communicating regard for the interviewee and her contribution. Themes and specific strategies that emerged in each of these thematic areas are summarized in Table 1 and described with illustrations below.

Table 1.

Themes related to effective interviewer strategies for collection of narrative data through telephone interviews

Overarching themes	Strategies
Cultivating rapport and maintaining connection	Informal conversational exchanges Orienting statements Sharing reciprocal information
Demonstrating responsiveness to interviewee content and concerns	Active listening Supportive vocalizations Validation and clarification
Communicating regard for the interviewee and her contribution	Acknowledgement of disclosure Statements of appreciation Respectful attention

Cultivating rapport and maintaining connection

This was critical to engaging interviewees and to maintaining a productive interviewer-interviewee relationship. Specific strategies identified in this general area included being friendly and personable through informal conversational exchanges, providing orienting statements to help guide the participants through the interview, and providing occasional reciprocal information.

Informal conversational exchanges

This included small talk before the start of the interview to put the interviewee at ease, friendly conversational exchanges, and the use of humor (light comments, laughter, and joking exchanges). Interviewers would sometimes participate in brief “side conversations,” then guide the respondent back to the interview, creating a balance between naturalness and structure. Some of the informal exchanges were directly related to the content of the interviews. For example, the following interchange occurred in the context of a question about a time that the respondent consumed significantly more alcohol that was typical for her (in this case, the respondent was with friends, who were driving, on her birthday):

Interviewee: …but typically I’m the driver, so I don’t drink, but since it was my birthday

Interviewer: But it was your birthday after all <both laugh>.

Interviewee: Right, it was my 30th birthday—so I was like, “Well, the hell with that.”

Orienting Statements

These statements were informally woven throughout the interview and commonly used to help participants know what to expect. Orienting statements early in the interview helped to inoculate the interviewee against feeling unheard when questions were redundant, illustrated by the following interviewer statement: “Okay, and I’m real sorry if this sounds redundant, it’s the nature of the beast here, I’m asking questions that may sound similar, but they’re kind of different, they have different nuances.” Later in the interview, orienting statements helped reinforce interviewee participation when interview fatigue might otherwise become an issue, such as: “So, we’re just about done and—again, I want to thank you for your patience—I’m going to ask a few final questions.” Orienting statements included giving permission to the interviewee to add information at any time: “At any point, if you want to add anything else that you think you want me to know, might be important to the study or anything, you’re certainly welcome to do that too, okay?” Sometimes statements reinforced interviewee rights to exercise control during the interview, illustrated by the exchange below:

Interviewer: Okay. I’m going to move onto the substance use theme, and, of course, you just talked to me rather extensively about your alcohol experience. I’m going to go ahead and ask the question.

Interviewee: Okay. <laughs> Am I allowed to say I already answered it?

Interviewer: You certainly are.

Interviewee: Okay. <both laugh> Well, maybe I’ll even expound.

Reciprocity

This involved interviewers giving back to interviewees by briefly sharing personal information pertinent to the topic in order to validate or encourage interviewees. For example, the human subjects protocol included procedures for checking in with clients about their level of distress; one interviewer made this check-in less intrusive through her personal disclosure: “I lost both my parents and I still get very emotional and so I’m not saying you’re doing anything wrong, it’s perfectly fine with me, I just want to make sure you’re okay.” Interviewers sometimes included informal comments about their own enthusiasm about the research project. In debriefing, one interviewer commented: “My excitement seemed to increase their excitement of being a part of such an important project.”

Demonstrating responsiveness to interviewee content and concerns

These strategies were also important to create a safe and empathetic interview environment. Specific strategies in this area included active listening, supportive vocalizations, and validation and clarification exchanges.

Active listening

This strategy included using of reflective and summary statements as well as follow-up questions specific to interviewee content. For example, the following reflective statement was followed by an interviewee-specific probe: “You mentioned it was about six months that you were really having a rough time. What did you do to cope during that period?” Although interviewers were not able to respond to visual cues, there were instances where it was possible to respond to both narrative content and tone of voice. For example, one interviewer noted, “I detect something in your tone,” which prompted the interviewee to elaborate in detail about conflicts with a family member.

Supportive vocalizations

These vocalizations included encouraging tones, encouraging words, and non-language encouragement such as the following: “sure,” “right,” “yeah,” “I know what you mean,” “mm-hmm,” “wow,” “okay” “I see,” and “interesting.”

Validation and clarification exchanges

These exchanges involved the interviewer checking to ensure accuracy and understanding, responding to interviewee requests for clarification about questions, and reassuring interviewees about the content and quality of their responses. The following exchange typified this interaction—Interviewee: “Am I being to lengthy?” Interviewer: “No, absolutely not. I’m enjoying listening to you.” Interviewees also expressed insecurity about whether the content of their narrative was valuable, such as, “I don’t know if that answers your question or not,” or “I don’t know if I’m answering this very well,” which prompted reassuring comments from interviewers such as, “There are no wrong answers you can give me” or “You’re doing just fine.”

Communicating regard for the interviewee and her contribution

This involved acknowledgement of disclosure and statements of appreciation. For example, interviewers frequently expressed appreciation for disclosure of personal information, particularly in response to sensitive questions, exemplified by the following exchange:

Interviewer: Thank you very much for sharing that with me. I know that’s very personal, and it can be hard to share with someone on the phone, a stranger, but I want to thank you for being open and sharing that information with me.

Interviewee: It’s really easier with a stranger than it is with someone that you know.

Interviewers also demonstrated respectful attention by maintaining an accepting, non-judgmental tone. Respect and positive regard were often communicated through simple affirming statements. This is illustrated by an exchange with one young sexual minority respondent during her response to questions about identity:

Interviewee: I mean, a couple of months ago I had pink hair in (names rural state). Nobody has pink hair in (state).

Interviewer: That was pretty bold, huh?

Interviewee: Exactly, but I liked it, and I do what I like….

It was notable that interviewees often expressed appreciation for “feeling heard” toward the end of the interview or otherwise commented about what they were getting out of participating in the study. For example, one interviewee commented, “Well, my experiences would not help anyone if I did not share them. And since you called and asked for it, I was more than glad to help you with it.” Upon being given gift certificate information at the end of the interview, another interviewee noted: “Oh yeah, I forgot all about that. That’s not why I was doing this—I would have done it anyway.”

Advantages, Challenges, and “Lessons Learned” for Addressing Challenges

The ability to conduct interviews across geographically disparate interviewees in the U.S. was a critical advantage for the purposes of our study. It would not have been feasible, in term of cost or logistics, to conduct follow-up interviews with a sub-sample of former respondents of a national sample through in-person interviews. The use of the telephone proved to maximize flexibility for interviewee scheduling and re-scheduling interviews. The telephone mode also made it possible for interviewees to schedule interviews at times that suited them and, in several cases, re-schedule. Some interviewees missed or were “late” to interview appointments, and research team members were able to reschedule. In addition to logistical advantages, it appeared that interviewees were comfortable creating privacy for their interview call and were open to providing details about their life experiences over the telephone. For example, one interviewee asked to skip a question and then, when the interviewer later asked if she was willing to answer the question (related to past trauma), the interviewee explained that her boyfriend had been in the room earlier and she postponed answering until he left.

The research team anticipated challenges in recruiting interviews given a lapse in time between the initial NAS interview and the follow-up calls (approximately 1–2 years) and the necessity for making “cold calls” in cases where mailing addresses were not available for sending advance notification. Other challenges included establishing rapport over the telephone and scheduling interviews in the context of a small research team with limited hours for scheduling interviews. Through debriefing, the team identified a number of “lessons learned” for addressing or minimizing challenges, which included the following: substantial pre-interview preparation, training, de-briefing, and effective use of relational skills.

Pre-interview preparation

Lessons learned from the debriefing include the following:

• In the absence of opportunities for face-to-face participant recruitment, options for communicating with participants in advance were critical. Specific strategies employed by the team and deemed helpful in retrospect included the following: developing telephone introductory and message scripts, sending introductory letters (by mail or email) to prospective interviewees for whom address information was available, establishing toll-free numbers for call-backs to minimize possible cost barriers for respondents returning calls from land-lines or from work, and creating a project web-page that prospective interviewees could visit to learn about the project and verify the legitimacy of the project,.

• Communication about designated blocks of time for interviews was more important than having mechanisms for access at any time. We initially put substantial effort into identifying a way to forward calls to project-specific cells phones to avoid losing opportunities to arrange interviews when respondents called back (as it was not feasible to support full time staffing). Interviewees frequently (and to the initial surprise of the researchers) agreed to be interviewed at the time that they answered the telephone. It is worth noting that the scripts for leaving telephone messages allowed the research team to mention the next windows in which calls would be made (typically on one or two weekday evenings and Saturdays). Participants may have already deliberated about their willingness to participate and, as such, simply answered the phone when they were inclined to be interviewed.

• Contacting former participants of a national population-based survey who are geographically dispersed also required pre-planning. For example, as part of a protocol for addressing possible distress during interviews, national and local referral resources (based on telephone area code or zip code if addresses were available) were compiled in advance of interviews. In addition, interviewers considered time zones when placing calls and scheduling interviews.

Training, debriefing and relational skills

We identified a number of other lessons learned in the process of implementing the research project that may be salient to individuals seeking to conduct in-depth interviews using population-based samples. These are outlined below:

• Selection of interviewers with strong interpersonal skills is an asset, particularly in the context of the need to establish rapport quickly and to conduct a semi-structure interview in a manner that is as “conversational” as possible. Interviewers also require skills in self-reflexivity, including the ability to reflect on the identities, social locations, assumptions, and the life experiences they bring to the research endeavor and their interactions with interviewees. Strong relational skills and competence in self-reflexivity helped to ensure that interviewers were authentic, attentive, able to critically examine their own reactions and responses, and open to working through tense or awkward moments that may arise in the interview interaction.

• Preparation for interviewers through training, mock interviews, and review of procedures in the event of interviewee distress is important to maximizing the opportunity to obtain rich, quality data. Orientation and procedures for using digital recorders designed for telephone interviews helped to avoid loss of data through technological errors.

• Chatting and small talk in advance of the interviews was critical to establishing rapport. Small talk often centered on weather, challenges of scheduling, or the research project. Friendly, clear, but not overly “canned” communications in advance of interviews was also helpful.

• Friendly persistence was helpful to reaching and retrieving prospective interviewees. For example, 8 of the 32 sexual minority women included in the final sample participated in interviews after multiple initial contact attempts and a gap of several (2–4) months. There were no significant differences in the number of contacts, or length of interviews, between respondents recruited using a “cold call” list (no addresses available for letters in advance of first call) and a list of respondents for whom advance information was available.

• De-briefing sessions after interviews are important for addressing the many issues that inevitably arise—from observations about what worked with interviewees to frustrations about rejection from prospective interviewees.

Discussion

Findings from this study identified specific strategies that “worked” in conducting qualitative interviews over the phone. Specific strategies for success in interviews emerged in three overarching areas: 1) cultivating rapport and maintaining connection, 2) demonstrating responsiveness to interviewee content, concerns, and 3) communicating regard for the interviewee and her contribution. Use of the telephone for qualitative interviews was advantageous for the purposes of conducting follow-up interviews with geographically dispersed former respondents of a national survey. Through debriefing, the team identified a number of “lessons learned” for addressing potential challenges, including the importance of substantial pre-interview preparation, training, effective use of relational skills, and post-interview reviews.

In general, our experience affirms the both the viability and value of telephone interviews as a method for collecting rich narrative data with women from marginalized populations, including data related to sensitive subjects. Although there was variability in interview length, we found that interviewees were generally willing to engage in lengthy interviews (average of one hour) over the telephone and that these interviews yielded rich data. This is consistent with other researchers who have similarly noted that telephone interviews yield high quality data (Lechuga, 2012; Holt, 2010; Saura and Balsas, 2014; Irvine, 2011; Sturges and Hanrahan, 2004; Sweet, 2002; Trier-Bieniek, 2012; Vogl, 2013; Musselwhite et al., 2007; Stephens, 2007; Cachia and Millward, 2011; Glogowska et al., 2011). Interviewees provided rich descriptions of their life experiences, including experiences related to sensitive topics, such as sexual identity, alcohol or other drug use, and traumatic experiences. Sexual minority women were notably open about their sexual identity, their current and past relationships, and, in many cases, about stressful experiences of discrimination or dynamics of acceptance and rejection in families of origin. In particular, our findings echo that of Trier-Bieniek, who suggests that telephone interviews are effective, and possibly preferable, when talking with interview participants about personal and sometimes painful experiences, because “they are being interviewed in familiar, comfortable settings and can dictate the course and direction of the interview” (p. 642). Furthermore, analyses of non-responders, comparisons of demographics between the in-depth interview sample and the original population-based NAS sample, and evaluation of response rates between prospective interviewees reached through “cold calls” and those receiving advance notice, all suggest that it is possible to obtain a sample of respondents for in-depth qualitative interviews that approximate the original probability sample.

Cachia and Millward (2011) observed that interviewees were likely to invest time in interviews “if they are sufficiently motivated and rapport has been successfully established” (p. 273). We found that small talk helped to establish rapport over the phone before the formal start of the interview and that continued affirming, encouraging exchanges helped to sustain a productive interview relationship. Participant motivation was often revealed at the end of the study (often after the close of the formal interview), and it was notable that many interviewees made explicit statements about feeling valued or heard by sharing their stories, finding meaning in contributing to a project that might ultimately benefit others, and even expressing ambivalence about accepting the cash incentive. Although motivation was not a focus of the study, our finding in this area is consistent with an observation by Cachia and Millard that the “positive impact of the cash incentive is reduced for people who are motivated by community involvement” (p. 268).

Interviewees frequently elicited reassurance or validation about their responses, which has been observed by other researchers (Holt, 2010; Irvine et al., 2013). Holt described “participant’s preoccupation about how they should perform” as participant anxiety, suggesting that this phenomena may be more apparent for some groups than others (p. 118). Future studies might examine more explicitly whether this dynamic is more typical of populations that have experienced marginalization. Similar to others (Holt, 2010; Saura and Balsas, 2014), we found that vocalizations, such as “uh-huh” and “right,” served as cues that the interviewer was listening in the absence of non-verbal cues, such as nods. Although less commonly mentioned in other studies, interviewers in our study often used orienting statements to guide respondents through the interview. The use of these kinds of orienting statements may have been necessitated in part by the follow-up sample (e.g., explaining that the kinds of questions in the qualitative interview would differ from their prior experience with a telephone survey) and by the inclusion of sensitive topics areas, such as substance use and past experiences of trauma.

This study has a number of limitations. Although the sample is drawn from former respondents of a population-based survey, it cannot be presumed that findings would be generalizable to the groups of women represented in the study. Although interviewees were similar in characteristics to prior NAS interviewees in many areas, they were generally higher in education and older than the NAS sample as a whole. Interviewees who were not available (e.g., who moved or whose numbers had changed) were likely different than those who still had the same active phone numbers at the time of the follow-up contact. Interview tape-recording began after an initial conversation with prospective interviewees and after conducting informed consent over the phone; consequently, transcriptions that would capture specific language and exchanges that occurred before taping were not available for analysis. Furthermore, the interview guide did not include questions specific to participant perceptions of their interview experience; consequently, it was not possible to describe what “worked” about interviews from the perspective of interviewees. Finally, metadata regarding number of visitors to the website was not collected and, other than a small number of contacts that were initiated through the website, it is not possible to document the degree to which the website was used or may have helped prospective participants become familiar and comfortable with the study before the telephone interview.

In spite of these limitations, findings of the study suggest that it is feasible to conduct in-depth interviews by telephone with a follow-up sample of respondents from a population-based survey, and to successfully engage sub-samples of women who are often considered marginalized or hard-to-reach. Findings underscore the opportunity afforded by telephone interviews for obtaining rich data, even when addressing sensitive topics such as substance use and past experiences of trauma.